Antimicrobial Chemotherapy, 4th Edition

Resistance to antimicrobial agents


The genetics of resistance

  1. J. Towner

All the properties of a microbial cell, including those of medical importance such as antibiotic resistance and virulence determinants, are determined ultimately by the microbial genome, which in turn comprises the three sources of genetic information in the cell: the chromosome, plasmids, and bacteriophages. Resistance of bacteria to antimicrobial agents may be either intrinsic or acquired (see Chapter 11). Intrinsic resistance is the ‘natural’ resistance possessed by a bacterial species and is usually specified by chromosomal genes. An example of a bacterial species with a high degree of intrinsic resistance is Pseudomonas aeruginosa. In contrast to intrinsic resistance, acquired resistance occurs in sensitive cells, either following alterations to the existing genome or by transfer of genetic information between cells. Thus a fundamental knowledge of microbial genetics is essential for an understanding of the development and spread of resistance to antimicrobial drugs.

The heritable information which specifies a bacterial cell, and which passes to daughter cells at cell division, is carried in bacteria, as in all living cells, as an ordered sequence of nucleotide pairs along molecules of DNA. The process of transcription of this information into messenger RNA, and its subsequent translation into functioning proteins by ribosomes, is also similar in bacteria and in other cells.

The bacterial chromosome

The main source of genetic information in a bacterial cell is the chromosome. Each bacterial cell has a single chromosome which, in the vast majority of cases, is known to form a single closed circular DNA molecule. In Escherichia coli, the organism studied most intensively, this single DNA molecule comprises about 4 × 103 kb (kilobases) and is about 1.4 mm in length.The chromosome is coiled and looped into a compact bundle in the cytoplasm of the cell, but it is not


separated from the cytoplasm by any form of nuclear membrane. Transcription of the DNA and translation of the resulting messenger RNA can therefore proceed simultaneously. Most bacterial chromosomes contain sufficient DNA to encode for 1000–3000 different genes. Not all of these genes are required to be expressed at any one time, and indeed it would be energetically wasteful for the cell to do so. Gene regulation is therefore necessary, and this can occur at either the transcriptional or translational level.

Chromosomal mutations to antibiotic resistance

Mutations result from rare mistakes in the DNA replication process and occur at the rate of between 10-4 and 10-10 per cell division. They usually involve deletion, substitution, or addition of one or only a few base pairs, which causes an alteration in the amino-acid composition of a particular peptide product. Such mistakes are random and spontaneous. They occur continously in all the genes of a cell and are independent of the presence or absence of a particular anti- biotic. The vast majority of mutations are repaired by the cell without any noticeable effect. The only influence that an antibiotic has is to select for the occasional spontaneous antibiotic-resistant mutant that may be present in a large sensitive population. In such a situation, the sensitive cells will be killed by the antibiotic whereas the resistant mutant will survive and proliferate to become the predominant type. Most chromosomal resistance mutations result in alter- ations to permeability or specific antibiotic target sites, but some result in enhanced production of an inactivating enzyme or bypass mechanism. The latter type are mutations at the transcriptional or translational level in gene regulatory mechanisms.

Chromosomal mutations to antibiotic resistance can be divided into single- step and multi-step types:

  • Single large-step mutations.With these mutations, a single mutational change results in a large increase in the minimum inhibitory concentration (MIC) of a particular antibiotic. Single-step mutations may lead to treatment failure when these drugs are used alone (seep. 181). In some Gram-negative bacilli, mutations in the genetic regulatory system for the normally low-level chromosomal β-lactamase may result in a vast overproduction of this enzyme and slow hydrolysis of compounds such as cefotaxime and ceftazidime that are considered under normal circumstances to be β-lactamase stable.
  • Multi-step mutations.These are sequential mutations that result in cumulative gradual stepwise increases in the MIC of a particular antibiotic. They are clinically quite common, especially in situations in which only low concentrations of antibiotic can be delivered to the site of an infection.




The bacterial chromosome carries all the genes necessary for the survival and replication of the bacterial cell under most circumstances. Many, perhaps all, bacteria also carry additional molecules of DNA (usually between 2 and 200 kb in size) known as plasmids, which are separate from, and normally replicate independently of, the bacterial chromosome. Plasmids can carry genes which confer a wide range of properties on the cells which carry them. In general, these are properties that are not essential for the survival of the cell under normal circumstances, but which offer the cells a survival advantage in unusual or adverse conditions. Examples of such properties are:

  • fertility: the ability to conjugate with and transfer genetic information into other bacteria (see later)
  • resistance to antibiotics: most antibiotic resistance encountered clinically is associated with plasmids
  • ability to produce bacteriocins: proteins inhibitory to other bacteria which may be ecological competitors toxin production
  • immunity to some bacteriophages.
  • ability to use unusual sugars and other substrates as foods.

Many different kinds of plasmid have been described. Plasmids differ in size, in DNA base composition, in the DNA fragments that can be recognized after treatment with restriction endonucleases (‘plasmid fingerprints’) and in their incompatibility behaviour. Compatible plasmids can coexist in the same host cell, while incompatible plasmids cannot, and so tend to be unstable and displace one another. There are at least 20 incompatibility (Inc) groups within the plas- mids found in Enterobacteriaceae, and similar incompatibility schemes are used to subdivide staphylococcal plasmids and those found in Pseudomonas spp.


The third possible source of genetic information in a bacterial cell is a bacterio- phage. Bacteriophages (phages) are viruses which infect bacteria. Most phages will attack only a relatively small number of strains of related bacteria—they have a narrow and specific host range. Phages can be divided into two main types:

  • Virulent phages inevitably destroy by lysis any bacteria that they infect, with the release of numerous new phage particles from each lysed cell.
  • Temperate (lysogenic) phages may either lyse or lysogenize infected bacterial cells. In the state of lysogeny, the phage nucleic acid is replicated in a


stable and dormant fashion within the infected cell, often following insertion into the host cell chromosome. Such a dormant phage is known as a prophage. However, while in the prophage state, some prophage genes may be expressed and may confer additional properties on the cell. Once in every few thousand cell divisions, a prophage becomes released from the dormant state and enters the lytic cycle, with subsequent destruction of its host cell and release of new phage particles into the surrounding medium.

Transfer of genetic information

There are three ways in which genetic information can be transferred from one bacterial cell into another: transformation, transduction, and conjugation.

  • Transformation involves lysis of a bacterial cell and the release of naked DNA into the surrounding medium. Under certain circumstances, some of this DNA can be acquired by intact bacterial cells in the vicinity. This process has been much studied in the laboratory, but there are few convincing demonstrations of its occurrence in vivo. The process depends crucially on the ability of the recipient cells to be competent for uptake of free DNA.
  • Transduction involves the accidental incorporation of bacterial DNA, either from the chromosome or a plasmid, into a bacteriophage particle during the phage lytic cycle. The phage particle then acts as a vector and transfers the bacterial DNA to the next cell which it infects.
  • Conjugation involves physical contact between two bacterial cells. The cells adhere to one another and DNA passes unidirectionally from one cell, termed the donor, into the other, the recipient. Ability to conjugate depends on carriage of an appropriate plasmid or transposon (see later) by the host cell.

The existence of these transfer mechanisms means that bacteria do not have to rely solely on a process of mutation and selection for their evolution. By such mechanisms they can acquire and express blocks of genetic information that have evolved and been refined elsewhere. A bacterial cell can, for example, acquire by conjugation a plasmid which carries genes conferring resistance to several different antibiotics, and as a result, within a very short time following the receipt of such a plasmid by a sensitive cell, the organisms in a given niche may change from a state in which they are all sensitive to a state in which it is very difficult to find an antibiotic that is effective against them. Of course, the ability to transfer genes in this way does not eliminate the need to evolve them in the first place but, once they have evolved, it ensures their eventual wide- spread dissemination under appropriate selection pressures.



Evolution of new resistance gene combinations

The distinction between chromosomal and plasmid genes is not absolute. Where appropriate regions of DNA homology exist, classic (‘normal’) recombination can occur, both between different plasmids and between plasmids and the chromosome. Although this process can lead to the formation of new antibiotic resistance gene combinations, it is relatively uncommon in bacteria because there are few regions of sequence homology between the bacterial chromosome and plasmids which can be exploited for this purpose. A more important mechanism by which antibiotic resistance genes can pass from one bacterial replicon to another is the ‘illegitimate’ recombination process known as transposition.


Transposition depends on the existence of specific genetic elements termed trans-posons. These elements are discrete sequences of DNA capable of translocation (transposition) from one replicon (plasmid or chromosome) to another. Unlike classic (‘normal’) recombination, transposons do not share extensive regions of homology with the replicon into which they insert. In many cases, transposons consist of individual resistance genes, or groups of genes, bounded by DNA sequences called either direct or inverted repeats—a sequence of bases at one end of the transposon which also appears, either in direct or reverse order, at the other end. These repeats may be relatively short, often of the order of 40 base pairs, but longer examples have been identified. It is likely that these DNA sequences provide highly specific recognition sites for certain enzymes (trans-posases) that catalyse the movement of transposons from one replicon to another without the need for extensive regions of sequence homology. Depending upon the transposon involved, insertion may occur at only a few or at many different sites on the host replicon. Transposons may carry genes conferring resistance to many different antibiotics, as well as other metabolic properties, and their existence helps to explain how a single antibiotic resistance gene can become disseminated over a wide range of unrelated replicons.

Isolated DNA sequences analogous to the terminal sequences of transposons can also move from one replicon to another, or be inserted in any region of any DNA molecule. Such insertion sequences contain no known genes unrelated to insertion functions but, in principle at least, two similar insertion sequences could bracket any assemblage of genes and convert it into a transposon. Thus, theoretically, all replicons are accessible to transposition and all genes are potentially transposable. This is of crucial evolutionary importance since it explains how genes of appropriate function can accumulate on a single replicon under the impact of selection pressure. Transposons and insertion sequences therefore play a vital part in plasmid evolution.




Transposons may contain combinations of genes conferring resistance to various different antibiotics. An important question concerns the mechanism by which new combinations of antibiotic resistance genes can be formed. It is now apparent that special molecular structures, termed integrons, may enable the formation of new combinations of resistance genes within a bacterial cell, either on a plasmid or within a transposon, in response to selection pressures.

Integrons appear to consist of two conserved segments of DNA located either side of inserted antibiotic resistance genes. Individual resistance genes seem to be capable of insertion or removal as ‘cassettes’ between these conserved structures. The cassettes can be found inserted in different orders and combinations. Integrons also act as an expression vector for ‘foreign’ antibiotic resistance genes by supplying a promoter for transcription of cassettes derived originally from completely unrelated organisms. Integrons lack many of the features associated with trans-posons, including direct or inverted repeats and functions required for transposition. They do, however, possess site-specific integration functions, notably a special enzyme termed an integrase. At least four different classes of integrons are known to exist and these are distinguished by the sequence of their integrase genes.

The precise role of integrons in the evolution and spread of antibiotic resistance genes remains to be determined, but they have been found, together with their associated antibiotic resistance gene cassettes, in many different Gram-negative bacteria. At least three potential mechanisms of spread exist:

  • the potential mobility of an integron itself by site-specific insertion
  • spread following insertion of an integron into a transposon
  • horizontal transfer of integrons on plasmids.

Whatever the mechanism, unrelated clinical isolates from different worldwide locations have been shown to carry the same integron structures, and it seems that these structures may play a key role in the formation and dissemination of new combinations of antibiotic resistance genes.

The process of evolution and spread of antibiotic resistance genes is still continuing. The origin of resistance genes carried by integrons, transposons, or plasmids, or even the origin of these elements themselves, is not known, but it has been possible to observe a steady increase in the numbers of resistant bacterial strains following the introduction of successive chemotherapeutic agents into clinical use. There are many examples and the evolutionary process is a continuous event.

Genotypic resistance

To summarize the earlier discussion, genes conferring resistance to antibiotics are often found inserted into integrons, and may be part of the bacterial chromosome


or may be carried on plasmids, transposons, or as part of a phage genome. The distribution of these genes between the chromosome and other elements reflects to some extent the biochemical mechanisms involved. For example, resistance which results from mutational alteration of an existing target protein will normally be chromosomal in location and will not be integron-associated, whereas resistance genes for entirely new enzymes, such as the aminoglycoside-modifying enzymes, novel β-lactamases, or trimethoprim-resistant dihydrofolate reductases, are commonly carried on plasmids and transposons as part of integrons. This reflects the fact that the evolution of any new enzyme is likely to be a very long process; the occurrence of the genes for such enzymes on plasmids, transposons, and integrons enables spread of these genes between different strains, species, and genera rather than requiring evolution of the genes afresh by each bacterial strain for itself.

Chromosomal and plasmid-mediated types of resistance may be equally important in the management of an individual patient. However, the plasmid-encoded variety has achieved greater notoriety because of the spectacular fashion in which bacteria may acquire resistance to a number of unrelated agents by a single genetic event. Certainly, it has been plasmid-encoded resistance which has caused most problems in the highly selective environment of the hospital. Nevertheless, mutational resistance involving the bacterial chromosome is a common cause of treatment failure with some compounds (p. 181) and it is strange but true that antibacterial agents for which resistance is not known to be encoded on plas-mids (e.g. quinolones and rifampicin) generally suffer from mutational resistance problems instead.

Phenotypic resistance

So far as is known, phenotypic resistance to antibacterial agents is rare, although it is not always possible to be sure that phenotypic changes brought about in the microenvironment of a lesion do not contribute to insusceptibility of bacteria in the infected host. In the laboratory, phenotypic resistance can sometimes be induced; for example, the outer envelope of Ps. aeruginosa can be altered by varying the conditions of growth, and this affects susceptibility to polymyxins.

Another example is the failure of penicillins and cephalosporins to kill persisters (those cells in a bacterial population that survive exposure to concentrations of β-lactam agents lethal to the rest of the culture). This does not result from a genetic event since the resistance is not heritable, and it is probable that the ‘resistant’ bacteria are caught in a particular metabolic state at the time of first encounter with the drug.

A peculiar form of phenotypic resistance is observed with mecillinam, a β-lactam antibiotic which, unusually, does not affect bacterial cell division. Mecillinam induces surface changes in susceptible Gram-negative bacilli which generally lead to death of the bacteria by osmotic rupture (p. 27). However,


those bacterial cells in the population that happen to have a low internal osmo-lality survive and, since mecillinam lacks the ability to prevent growth and division, such bacteria continue to grow in a morphologically altered form. On withdrawal of the drug, the bacteria resume their normal shape and, in due course, revert to the same mixed susceptibility as the original parent culture.

The influence of antibiotic selection pressure

Antibiotic resistance genes, and the genetic elements which carry them, existed before the introduction of antibiotics into human medicine. It has become established beyond doubt that the emergence and survival of predominantly resistant bacterial populations is a result of the selective pressure associated with the widespread use of antibiotics. Resistant cells survive in a given niche at the expense of sensitive cells of the same or other species. However, drug resistance, whether chromosomal or plasmid-encoded, often carries metabolic penalties for the cell. If the selective pressures of antibiotic use are withdrawn, there is sometimes a gradual return to sensitivity in the population. Individual cells may lose their plasmids and chromosomal mutations may revert to sensitivity. The implications of this process for efforts to control and limit the spread of bacterial drug resistance are discussed in the next chapter.